Impact of Professional Background on Inter-Annotator Variability and Accuracy During Annotation of Clinical Notes.

BACKGROUND: The aging population's need for treatment of chronic diseases is exhibiting a marked increase in urgency, with heart failure being one of the most severe diseases in this regard. To improve outpatient care of these patients and reduce hospitalization rates, the telemedical disease management program HerzMobil was developed in the past. OBJECTIVE: This work aims to analyze the inter-annotator variability among two professional groups (healthcare and engineering) involved in this program's annotation process of free-text clinical notes using categories. METHODS: A dataset of 1,300 text snippets was annotated by 13 annotators with different backgrounds. Inter-annotator variability and accuracy were evaluated using the F1-score and analyzed for differences between categories, annotators, and their professional backgrounds. RESULTS: The results show a significant difference between note categories concerning inter-annotator variability (p<0.0001) and accuracy (p<0.0001). However, there was no statistically significant difference between the two annotator groups, neither concerning inter-annotator variability (p=0.15) nor accuracy (p=0.84). CONCLUSION: Professional background had no significant impact on the annotation of free-text HerzMobil notes.

Formalizing the Austrian Procedure Catalogue: A 4-step methodological analysis approach.

OBJECTIVES: Due to the lack of an internationally accepted and adopted standard for coding health interventions, Austria has established its own country-specific procedure classification system - the Austrian Procedure Catalogue (APC). Even though the APC is an elaborate coding standard for medical procedures, it has shortcomings that limit its usability. In order to enhance usability and usefulness, especially for research purposes and e-health applications, we developed an ontologized version of the APC. In this paper we present a novel four-step approach for the ontology engineering process, which enables accurate extraction of relevant concepts for medical ontologies from written text. METHODS: The proposed approach for formalizing the APC consists of the following four steps: (1) comparative pre-analysis, (2) definition analysis, (3) typological analysis, and (4) ontology implementation. The first step contained a comparison of the APC to other well-established or elaborate health intervention coding systems in order to identify strengths and weaknesses of the APC. In the second step, a list of definitions of medical terminology used in the APC was obtained. This list of definitions was used as input for Step 3, in which we identified the most important concepts to describe medical procedures using the qualitative typological analysis approach. The definition analysis as well as the typological analysis are well-known and effective methods used in social sciences, but not commonly employed in the computer science or ontology engineering domain. Finally, this list of concepts was used in Step 4 to formalize the APC. RESULTS: The pre-analysis highlighted the major shortcomings of the APC, such as the lack of formal definition, leading to implicitly available, but not directly accessible information (hidden data), or the poor procedural type classification. After performing the definition and subsequent typological analyses, we were able to identify the following main characteristics of health interventions: (1) Procedural type, (2) Anatomical site, (3) Medical device, (4) Pathology, (5) Access, (6) Body system, (7) Population, (8) Aim, (9) Discipline, (10) Technique, and (11) Body Function. These main characteristics were taken as input of classes for the formalization of the APC. We were also able to identify relevant relations between classes. CONCLUSIONS: The proposed four-step approach for formalizing the APC provides a novel, systematically developed, strong framework to semantically enrich procedure classifications. Although this methodology was designed to address the particularities of the APC, the included methods are based on generic analysis tasks, and therefore can be re-used to provide a systematic representation of other procedure catalogs or classification systems and hence contribute towards a universal alignment of such representations, if desired.

Morphological changes in the human cervical intervertebral disc post trauma: response to fracture-type and degeneration grade over time.

PURPOSE: In the first 24 h post-intervertebral disc (IVD) trauma, up to 75 % cell death has been reported. In addition, burst fractures cause post-traumatic disc degeneration by elevated pro-apoptotic and pro-inflammatory gene transcription. Moreover, some patients have pre-trauma degenerative disc disease. The aim of the study was to assess histological changes and cell-death over a time period of up to 1 year caused by mechanical and structural factors. METHODS: 116 anterior portions of IVDs of the cervical spine were studied histologically by light microscopy and ultrastructurally by transmission electron microscopy (TEM). The group was investigated with regard to three main parameters: fracture mechanism (compressive vs. tensile/shear loads), degeneration grade (low vs. high) and endplate fracture (with vs. without). Disc architecture (e.g. ruptures) was studied histologically. Cell morphology was examined ultrastructurally to quantify cell-death, healthy and balloon cells. According to ultrastructural observations, two time-groups (up to 6 days vs. later) were established. Statistical analyses were carried out within and between time-groups. RESULTS: Histological changes were obvious in the annulus fibrosus where ruptures with haematoma were replaced by granulation tissue. Significant differences in cell-death were seen in the first few days due to different loads. In contrast to the more degenerated segments, low degenerated ones revealed significantly less cell death with time post-trauma. Interestingly, no difference was found between groups after the sixth day. Cell-death (mean 44 % for all investigated groups) remained high after day 6 post-trauma. CONCLUSION: IVDs retrieved from low grade degenerated segments revealed a significant recovery, with less cell-death and a partially restored disc matrix, although cell-death remained high. Long-term clinical studies of stabilized segments arising from different fracture mechanisms are required.

Morphological similarities after compression trauma of bovine and human intervertebral discs: Do disc cells have a chance of surviving?

To study the behavior of bovine disc cells and changes in disc matrix following in vitro compression tests; to compare the findings to investigations on human intervertebral discs (IVD) after burst fracture of the cervical spine. Healthy IVDs (n = 21) from three bovine tails were studied at 6 and 12 h post-mortem, with 16 IVDs subjected to impact loading and five as unloaded controls. IVDs (n = 8) from patients with burst fractures were compared to the bovine compression group. Specimens were studied macroscopically, histologically, and ultrastructurally for healthy cells, balloon cells, and disc cell death (DCD). Annulus ruptures were seen in both post-trauma groups, with radial ruptures being present histologically in all loaded bovine discs. Balloon cells were found in some human IVDs and were induced in vitro in bovine loaded discs within a distinct range of absorbed energy. There was a positive correlation between DCD and absorbed energy in all compartments of bovine discs. Both species showed similar patterns of DCD in the different compartments. This study was able to show similarities between both species in cell morphologies and matrix damage. The survival of the disc after substantial compression trauma thus seems to remain highly questionable.